Skip to content

Latest commit

 

History

History
185 lines (164 loc) · 6.29 KB

tutorial.md

File metadata and controls

185 lines (164 loc) · 6.29 KB
Raw

Tutorial

Submitting a folder of .traj files for optimization

Create a folder of folders containing the .traj files that you wish to optimize. For example

tree 
sn_bea_tmpo_structures
├── 00_closed
│   ├── 01SnBea
│   │   └── bea_structure_unopt.traj
│   ├── 02SnBea
│   │   └── bea_structure_unopt.traj
│   ├── 03SnBea
│   │   └── bea_structure_unopt.traj
│   ├── 04SnBea
│   │   └── bea_structure_unopt.traj
│   ├── 07SnBea
│   │   └── bea_structure_unopt.traj
│   ├── 08SnBea
│   │   └── bea_structure_unopt.traj
│   └── 09SnBea
│       └── bea_structure_unopt.traj
└── 02_open2
    ├── 01_Sn
    │   └── bea_structure_unopt.traj
    ├── 02_Sn
    │   └── bea_structure_unopt.traj
    ├── 03_Sn
    │   └── bea_structure_unopt.traj
    ├── 04_Sn
    │   └── bea_structure_unopt.traj
    ├── 07_Sn
    │   └── bea_structure_unopt.traj
    ├── 08_Sn
    │   └── bea_structure_unopt.traj
    ├── 09_Sn
    │   └── bea_structure_unopt.traj
    └── bea_structure_unopt.traj

Next determine the workflow for the optimization

Here is an example workflow

  1. Pre-opt bare zeolite – Single point for each ENCUT iterations, ENCUT (300, 400, …., 700)
  2. 1Al, 2Al zeolites – unconstrained opt, 50~100 steps for each ENCUT iterations, ENCUT (x, x, 700)
  3. EF/BronstedH/Defect zeolites – constrained opt, pre-opt with FF, 30~50 steps for each ENCUT iterations, ENCUT (x, x, 700)

Create a spec dictionary for each step of this process

Here is an example spec

spec1 = {"is_zeolite": True,  
        "database_path": db_url,  
        "input_id": 1,  
        "nsw": 1,  
        "my_nsw": 1,  
        "encut": 520.0,  
        "kpts": (1, 1, 1),  
        "ivdw": 12,  
        "isif": 2}

Use vasp-fw to submit all files for the same optimization

This is an example script makes the same workflow for all of the traj files in the folder, and submits the workflow to the fireworks launchpad.

import os  
import yaml  
from ase.io import read, write  
from vaspfw.optimize import OptimizeWithVaps  

with open(os.path.join('data', 'aws.yaml')) as f:  
    db_yaml = yaml.load(f, Loader=yaml.FullLoader)  

db_url = db_yaml['db']  
launchpad_path = os.path.join(os.getcwd(), 'data', 'my_launchpad.yaml')  
folder_path = "sn_bea_tmpo_structures" 

spec1 = {"is_zeolite": True,  
        "database_path": db_url,  
        "input_id": 1,  
        "nsw": 1,  
        "my_nsw": 1,  
        "encut": 400.0,  
        "kpts": (1, 1, 1),  
        "ivdw": 12,  
        "isif": 2}  

spec2 = {"is_zeolite": True,  
        "database_path": db_url,  
        "input_id": 1,  
        "nsw": 1,  
        "my_nsw": 1,  
        "encut": 600.0,  
        "kpts": (1, 1, 1),  
        "ivdw": 12,  
        "isif": 2}  

spec3 = {"is_zeolite": True, 
        "database_path": db_url,  
        "input_id": 1,  
        "nsw": 1,  
        "my_nsw": 1,  
        "encut": 800.0,  
        "kpts": (1, 1, 1),  
        "ivdw": 12,  
        "isif": 2}  

specs = [spec1, spec2, spec3]
opt_vasp = OptimizeWithVaps(folder_path=folder_path,  
                            file_format="traj",  
                            db_path=db_url,  
                            specs=specs,  
                            launchpad_path=launchpad_path,  
                            reset_launchpad=True)  

opt_vasp.submit()

The jobs have now been submitted to fireworks. Now you can go to your remote computer and use qlaunch to submit all of your jobs.

Once these jobs are done, you can now put the optimized structures into your original folder structure. To do this use the SaveStructures class in the save_results file.

import os  
from vaspfw.save_structures import SaveStructures  
import yaml  

with open(os.path.join('data', 'aws.yaml')) as f:  
    db_yaml = yaml.load(f, Loader=yaml.FullLoader)  

db_url = db_yaml['db']  
launchpad_path = os.path.join(os.getcwd(), 'data', 'my_launchpad_backup.yaml')  

save_structures = SaveStructures(db_path=db_url, launchpad_path=launchpad_path)  
save_structures.save()

This function prints out confirmations for each successful save and failure, but does not throw exceptions. After running this file check out your original folder structure. You should see an additional, optimized file in each folder.

├── sn_bea_tmpo_structures
│   ├── 00_closed
│   │   ├── 01SnBea
│   │   │   ├── opt_from_vasp.traj
│   │   │   └── optimized.traj
│   │   ├── 02SnBea
│   │   │   ├── opt_from_vasp.traj
│   │   │   └── optimized.traj
│   │   ├── 03SnBea
│   │   │   ├── opt_from_vasp.traj
│   │   │   └── optimized.traj
│   │   ├── 04SnBea
│   │   │   ├── opt_from_vasp.traj
│   │   │   └── optimized.traj
│   │   ├── 07SnBea
│   │   │   ├── opt_from_vasp.traj
│   │   │   └── optimized.traj
│   │   ├── 08SnBea
│   │   │   ├── opt_from_vasp.traj
│   │   │   └── optimized.traj
│   │   └── 09SnBea
│   │       ├── opt_from_vasp.traj
│   │       └── optimized.traj
│   └── 02_open2
│       ├── 01_Sn
│       │   ├── opt_from_vasp.traj
│       │   └── optimized.traj
│       ├── 02_Sn
│       │   ├── opt_from_vasp.traj
│       │   └── optimized.traj
│       ├── 03_Sn
│       │   └── opt_from_vasp.traj
│       ├── 04_Sn
│       │   ├── opt_from_vasp.traj
│       │   └── optimized.traj
│       ├── 07_Sn
│       │   ├── opt_from_vasp.traj
│       │   └── optimized.traj
│       ├── 08_Sn
│       │   ├── opt_from_vasp.traj
│       │   └── optimized.traj
│       ├── 09_Sn
│       │   ├── opt_from_vasp.traj
│       │   └── optimized.traj
│       ├── opt_from_vasp.traj
│       └── optimized.traj